Xuemeng Li

Dalian Minzu University, Minzu University of China

Papers

Total Citations

H-Index

About

Xuemeng Li is a researcher specializing in computer vision and deep learning, with a particular focus on object detection and feature fusion architectures. Their work addresses a fundamental challenge in deep learning-based detection systems: the loss of high-frequency detail and texture information during the feature extraction and up-sampling process common in modern neural networks. Li's most notable contributions center on innovative network architectures that intelligently integrate multi-scale feature representations. Their 2021 paper on the Convolutional Feature Frequency Adaptive Fusion Object Detection Network, which has garnered 9 citations, demonstrates a sophisticated approach to combining frequency-domain information within convolutional pipelines. Complementing this, their Pyramid Frequency Feature Fusion Object Detection Networks introduces a three-branch pyramid architecture specifically designed to recover fine-grained spatial details that are typically sacrificed in deep feature hierarchies. Both contributions emerged in 2021, reflecting Li's concentrated and productive research efforts in addressing real limitations of standard detection frameworks such as those built on Feature Pyramid Networks. Their work appeals to researchers seeking more faithful multi-scale representations, particularly in applications where fine texture and edge detail are critical to detection accuracy. With a growing citation record, Li represents an emerging voice in frequency-aware deep learning research.